1 
NOTE: GLK's approximate ranking of 5 most important tagged with 
NOTE: GLK's approximate ranking of 8 most important tagged with 
2 
[GLK:1], [GLK:2], ... 
[GLK:1], [GLK:2], ... 
3 


4 
============================== 
======================== 
5 
other SHORT TERM ============= (including needed for LIC) 
SHORT TERM ============= (*needed* for streamlines & tractography) 
6 
============================== 
======================== 




Add a clamp function, which takes three arguments; either three scalars: 


clamp(x, minval, maxval) = max(minval, min(maxval, x)) 


or three vectors of the same size: 


clamp([x,y], minvec, maxvec) = [max(minvec[0], min(maxvec[0], x)), 


max(minvec[1], min(maxvec[1], y))] 


This would be useful in many current Diderot programs. 


One question: clamp(x, minval, maxval) is the argument order 


used in OpenCL and other places, but clamp(minval, maxval, x) 


would be more consistent with lerp(minout, maxout, x). 

7 


8 
Level of differentiability in field type should be statement about how 
[GLK:3] Add sequence types (needed for evals & evecs) 

much differentiation the program *needs*, rather than what the kernel 


*provides*. The needed differentiability can be less than or equal to 


the provided differentiability. 





[GLK:1] Add sequence types (needed for evals & evecs) 

9 
syntax 
syntax 
10 
types: ty '{' INT '}' 
types: ty '{' INT '}' 
11 
value construction: '{' e1 ',' … ',' en '}' 
value construction: '{' e1 ',' … ',' en '}' 
12 
indexing: e '{' e '}' 
indexing: e '{' e '}' 
13 


14 
IL support for higherorder tensor values (matrices, etc). 
[GLK:4] evals & evecs for symmetric tensor[2,2] and 
15 
tensor construction [DONE] 
tensor[3,3] (requires sequences) 

tensor indexing [DONE] 


tensor slicing 


verify that hessians work correctly [DONE] 

16 


17 
Use ∇⊗ etc. syntax 
ability to emit/track/record variables into dynamically resized 
18 
syntax [DONE] 
runtime buffer 

typechecking 


IL and codegen 

19 


20 
test/uninit.diderot: 
tensor fields: convolution on general tensor images 

documents need for better compiler error messages when output variables 


are not initialized; the current messages are very cryptic 

21 


22 
determinant ("det") for tensor[3,3] 
======================== 
23 

SHORTISH TERM ========= (to make using Diderot less annoying to 
24 

======================== program in, and slow to execute) 
25 


26 
expand trace in mid to low translation 
valuenumbering optimization [DONE, but needs more testing] 
27 


28 
valuenumbering optimization 
Allow ".ddro" file extensions in addition to ".diderot" 
29 


30 
Add type aliases for color types 
Be able to output values of type tensor[2,2] and tensor[3,3] 
31 
rgb = real{3} 
(currently only scalars & vectors) 

rgba = real{4} 

32 


33 
============================== 
[GLK:1] Add a clamp function, which takes three arguments; either 
34 
MEDIUM TERM ================== (including needed for streamlines & tractography) 
three scalars: 
35 
============================== 
clamp(lo, hi, x) = max(lo, min(hi, x)) 
36 

or three vectors of the same size: 
37 

clamp(lo, hi, [x,y]) = [max(lo[0], min(hi[0], x)), 
38 

max(lo[1], min(hi[1], y))] 
39 

This would be useful in many current Diderot programs. 
40 

One question: clamp(x, lo, hi) is the argument order used in OpenCL 
41 

and other places, but clamp(lo, hi, x) is much more consistent with 
42 

lerp(lo, hi, x), hence GLK's preference 
43 


44 
[GLK:1] evals & evecs for symmetric tensor[3,3] (requires sequences) 
[GLK:2] Proper handling of stabilize method 
45 


46 
[GLK:2] Save Diderot output to nrrd, instead of "mip.txt" 
allow "*" to represent "modulate": percomponent multiplication of 
47 

vectors, and vectors only (not tensors of order 2 or higher). Once 
48 

sequences are implemented this should be removed: the operation is not 
49 

invariant WRT basis so it is not a legit vector computation. 
50 


51 

implicit type promotion of integers to reals where reals are 
52 

required (e.g. not exponentiation "^") 
53 


54 

[GLK:5] Save Diderot output to nrrd, instead of "mip.txt" 
55 
For grid of strands, save to similarlyshaped array 
For grid of strands, save to similarlyshaped array 
56 
For list of strands, save to long 1D (or 2D for nonscalar output) list 
For list of strands, save to long 1D (or 2D for nonscalar output) list 
57 
For ragged things (like tractography output), will need to save both 
For ragged things (like tractography output), will need to save both 
58 
complete list of values, as well as list of start indices and lengths 
complete list of values, as well as list of start indices and lengths 
59 
to index into complete list 
to index into complete list 
60 


61 
[GLK:3] Use of Teem's "hest" commandline parser for getting 
[GLK:6] Use of Teem's "hest" commandline parser for getting 
62 
any input variables that are not defined in the source file 
any input variables that are not defined in the source file 
63 


64 
[GLK:4] ability to declare a field so that probe positions are 
[GLK:7] ability to declare a field so that probe positions are 
65 
*always* "inside"; with various ways of mapping the known image values 
*always* "inside"; with various ways of mapping the known image values 
66 
to nonexistant index locations. One possible syntax emphasizes that 
to nonexistant index locations. One possible syntax emphasizes that 
67 
there is a index mapping function that logically precedes convolution: 
there is a index mapping function that logically precedes convolution: 
70 
F = bspln3 ⊛ (img ◦ mirror) 
F = bspln3 ⊛ (img ◦ mirror) 
71 
where "◦" or "∘" is used to indicate function composition 
where "◦" or "∘" is used to indicate function composition 
72 


73 
extend norm (exp) to all tensor types [DONE for vectors and matrices] 
Level of differentiability in field type should be statement about how 
74 

much differentiation the program *needs*, rather than what the kernel 
75 
ability to emit/track/record variables into dynamically resized 
*provides*. The needed differentiability can be less than or equal to 
76 
runtime buffer 
the provided differentiability. 




Want: allow X *= Y, X /= Y, X += Y, X = Y to mean what they do in C, 


provided that X*Y, X/Y, X+Y, XY are already supported. 


Nearly every Diderot program would be simplified by this. 

77 


78 
Want: nontrivial field expressions & functions: 
Use ∇⊗ etc. syntax 
79 
image(2)[2] Vimg = load(...); 
syntax [DONE] 
80 
field#0(2)[] Vlen = Vimg ⊛ bspln3; 
typechecking 
81 
to get a scalar field of vector length, or 
IL and codegen 

field#2(2)[] F = Fimg ⊛ bspln3; 


field#0(2)[] Gmag = ∇F; 


to get a scalar field of gradient magnitude, or 


field#2(2)[] F = Fimg ⊛ bspln3; 


field#0(2)[] Gmsq = ∇F•∇F; 


to get a scalar field of squared gradient magnitude, which is simpler 


to differentiate. However, there is value in having these, even if 


the differentiation of them is not supported (hence the indication 


of "field#0" for these above) 





Want: ability to apply "normalize" to a field itself, e.g. 


field#0(2)[2] V = normalize(Vimg ⊛ ctmr); 


so that V(x) = normalize((Vimg ⊛ ctmr)(x)). 


Having this would simplify expression of standard LIC method, and 


would also help express other vector field expressions that arise 


in vector field feature exraction. 

82 


83 
tensor fields: convolution on general tensor images 
Add type aliases for color types 
84 

rgb = real{3} 
85 

rgba = real{4} 
86 


87 
============================== 
============================== 
88 
other MEDIUM TERM ============ (needed for particles) 
MEDIUM TERM ================== (*needed* for particles) 
89 
============================== 
============================== 
90 



Put small 1D and 2D fields, when reconstructed specifically by tent 


and when differentiation is not needed, into faster texture buffers. 


test/illustvr.diderot is good example of program that uses multiple 


such 1D fields basically as lookuptablebased function evaluation 




91 
runtime birth of strands 
runtime birth of strands 
92 


93 
"initially" supports lists 
"initially" supports lists 
95 
"initially" supports lists of positions output from 
"initially" supports lists of positions output from 
96 
different initalization Diderot program 
different initalization Diderot program 
97 


98 
spatial data structure that permits strands' queries of neighbors 
Communication between strands: they have to be able to learn each 
99 

other's state (at the previous iteration). Early version of this can 
100 

have the network of neighbors be completely static (for running one 
101 

strand/pixel image computations). Later version with strands moving 
102 

through the domain will require some spatial data structure to 
103 

optimize discovery of neighbors. 
104 


105 

============================ 
106 

MEDIUMISH TERM ============ (to make Diderot more useful/effective) 
107 

============================ 
108 


109 
proper handling of stabilize method 
Python/ctypes interface to runtime 
110 


111 
test/vrkcomp2.diderot: Add support for code like 
support for Python interop and GUI 
112 


113 
(F1 if x else F2)@pos 
Allow integer exponentiation ("^2") to apply to square matrices, 
114 

to represent repeated matrix multiplication 
115 


116 
This will require duplication of the continuation of the conditional 
Alow X *= Y, X /= Y, X += Y, X = Y to mean what they do in C, 
117 
(but we should only duplicate over the liverange of the result of the 
provided that X*Y, X/Y, X+Y, XY are already supported. 
118 
conditional. 
Nearly every Diderot program would be simplified by this. 
119 


120 

Put small 1D and 2D fields, when reconstructed specifically by tent 
121 

and when differentiation is not needed, into faster texture buffers. 
122 

test/illustvr.diderot is good example of program that uses multiple 
123 

such 1D fields basically as lookuptablebased function evaluation 
124 


125 

expand trace in mid to low translation 
126 


127 

extend norm (exp) to all tensor types [DONE for vectors and matrices] 
128 


129 

determinant ("det") for tensor[3,3] 
130 


131 
add ":" for tensor dot product (contracts out two indices 
add ":" for tensor dot product (contracts out two indices 
132 
instead of one like •), valid for all pairs of tensors with 
instead of one like •), valid for all pairs of tensors with 
133 
at least two indices 
at least two indices 
134 


135 
============================== 
test/uninit.diderot: 
136 
other MEDIUM TERM ============ 
documents need for better compiler error messages when output variables 
137 
============================== 
are not initialized; the current messages are very cryptic 
138 


139 
want: warnings when "D" (reserved for differentiation) is declared as 
want: warnings when "D" (reserved for differentiation) is declared as 
140 
a variable name (get confusing error messages now) 
a variable name (get confusing error messages now) 
141 



support for Python interop and GUI 





Python/ctypes interface to runtime 





============================== 


LONG TERM ==================== 

142 
============================== 
============================== 
143 

LONG TERM ==================== (make Diderot more interesting/attractive from 
144 

============================== a research standpoint) 
145 


146 

IL support for higherorder tensor values (matrices, etc). 
147 

tensor construction [DONE] 
148 

tensor indexing [DONE] 
149 

tensor slicing 
150 

verify that hessians work correctly [DONE] 
151 


152 
Better handling of variables that determines the scope of a variable 
Better handling of variables that determines the scope of a variable 
153 
based on its actual use, instead of where the user defined it. So, 
based on its actual use, instead of where the user defined it. So, 
155 
scope. Also prune out useless variables, which should include field 
scope. Also prune out useless variables, which should include field 
156 
variables after the translation to midil. 
variables after the translation to midil. 
157 


158 

test/vrkcomp2.diderot: Add support for code like 
159 

(F1 if x else F2)@pos 
160 

This will require duplication of the continuation of the conditional 
161 

(but we should only duplicate over the liverange of the result of the 
162 

conditional. 
163 


164 

[GLK:8] Want: nontrivial field expressions & functions. 
165 

scalar fields from scalar fields F and G: 
166 

field#0(2)[] X = (sin(F) + 1.0)/2; 
167 

field#0(2)[] X = F*G; 
168 

scalar field of vector field magnitude: 
169 

image(2)[2] Vimg = load(...); 
170 

field#0(2)[] Vlen = Vimg ⊛ bspln3; 
171 

field of normalized vectors (for LIC and vector field feature extraction) 
172 

field#2(2)[2] F = ... 
173 

field#0(2)[2] V = normalize(F); 
174 

scalar field of gradient magnitude (for edge detection)) 
175 

field#2(2)[] F = Fimg ⊛ bspln3; 
176 

field#0(2)[] Gmag = ∇F; 
177 

scalar field of squared gradient magnitude (simpler to differentiate): 
178 

field#2(2)[] F = Fimg ⊛ bspln3; 
179 

field#0(2)[] Gmsq = ∇F•∇F; 
180 

There is value in having these, even if the differentiation of them is 
181 

not supported (hence the indication of "field#0" for these above) 
182 


183 

Introduce region types (syntax region(d), where d is the dimension of the 
184 

region. One useful operator would be 
185 

dom : field#k(d)[s] > region(d) 
186 

Then the inside test could be written as 
187 

pos ∈ dom(F) 
188 

We could further extend this approach to allow geometric definitions of 
189 

regions. It might also be useful to do inside tests in world space, 
190 

instead of image space. 
191 


192 
co vs contra index distinction 
co vs contra index distinction 
193 


194 
some indication of tensor symmetry 
Permit field composition: 
195 

field#2(3)[3] warp = bspln3 ⊛ warpData; 
196 

field#2(3)[] F = bspln3 ⊛ img; 
197 

field#2(3)[] Fwarp = F ◦ warp; 
198 

So Fwarp(x) = F(warp(X)). Chain rule can be used for differentation. 
199 

This will be instrumental for expressing nonrigid registration 
200 

methods (but those will require covscontra index distinction) 
201 


202 

Allow the convolution to be specified either as a single 1D kernel 
203 

(as we have it now): 
204 

field#2(3)[] F = bspln3 ⊛ img; 
205 

or, as a tensor product of kernels, one for each axis, e.g. 
206 

field#0(3)[] F = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img; 
207 

This is especially important for things like timevarying data, or 
208 

other multidimensional fields where one axis of the domain is very 
209 

different from the rest, and hence must be treated separately when 
210 

it comes to convolution. What is very unclear is how, in such cases, 
211 

we should notate the gradient, when we only want to differentiate with 
212 

respect to some subset of the axes. One ambitious idea would be: 
213 

field#0(3)[] Ft = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img; // 2D timevarying field 
214 

field#0(2)[] F = lambda([x,y], Ft([x,y,42.0])) // restriction to time=42.0 
215 

vec2 grad = ∇F([x,y]); // 2D gradient 
216 


217 

representation of tensor symmetry 
218 
(have to identify the group of index permutations that are symmetries) 
(have to identify the group of index permutations that are symmetries) 
219 


220 
dot works on all tensors 
dot works on all tensors 